Strange cosmic particles in my detector

[LuckyFox]

Hello! This is my first post on this forum. I am writing because I would be very interested to know the opinion of the forum participants on this issue. Almost ten years ago, I wrote a program Cosmic Ray Finder to record particle traces on a webcam (analogue of the project DECO) (the webcam is placed in an opaque container, the frame from the camera is read in a loop and it is determined whether there is a noticeable excess above the average brightness level - if so, the frame is saved).

In most cases, events are small dots, dashes, and fragments of curves:

But once on average for every hundred or two events, points of sharply increased size occur. Most often, several such events occur in one day, and then again there is a long break. And the strangest thing is that neighboring large dots (which can be separated in time by hours) can fall next to each other, even overlap (although the sizes of the matrix are several millimeters!).
Recent examples of "big spots":
two particles, separated in time by 136 minutes, actually hit the same point in the matrix:

group of three particles:

How can this be?!!!! Do the particles fly in a very narrow beam? Where?

 

[ohwilleke]

High energy cosmic ray particles aren't all that uncommon, and the far away events that produce them can produce them in bursts. If a far away event produces of burst of them, all of them that reach Earth will be in a line until they are diverted by collisions with atoms in the atmosphere (if they encounter any). See generally https://en.wikipedia.org/wiki/Cosmic_ray

 

[LuckyFox]

The strange thing about these events is not the high particle energy.
Such events most often occur in one day, followed by a long break.
Often such events are recorded one after another.
Often neighboring events fall into the same area of the camera, sometimes even overlapping. But the size of the camera matrix is several millimeters.
Imagine the probability of such a coincidence. And they happen more than once.

 

[Vanadium 50]

and the far away events that produce them can produce them in bursts.

But not on the timescale of a day.

This is almost surely a detector artifact of one sort or another. Just because the app says "cosmic rays" doesn't mean that's what they are. You are describing some kind of non-Poisson process, which is much more likely to be in your detector/readout chain than be from cosmics.

 

[Ibix]

If a far away event produces of burst of them, all of them that reach Earth will be in a line until they are diverted by collisions with atoms in the atmosphere (if they encounter any).

I don't think this is correct. I would expect events scattered all over the detector at more or less the same time from a short-lived high-energy event. Otherwise only one telescope would be able to observe them at once

I am writing because I would be very interested to know the opinion of the forum participants on this issue.

I have several questions. First is how sure are you that these events are actually cosmic rays? The DECO app claims to do some analysis of candidate events before deciding they are. Do you do more than naive thresholding? Have you tested it at different altitudes, for example? Or investigated what is needed to stop muons and tried shielding it?

Second, coincidences are more common than people think. Divide your detector up into a 20×20 grid - that's 400 areas. If you have two events drawn from a uniform distribution the chances are ##1-\frac{399}{400}## that they're in the same area. Three events and the chances are ##1-\frac{399\times 398}{400\times 400}## that at least two are in the same area. You actually only need 24 events before you've got better than a 50/50 chance that two occur in the same region. Google "birthday problem" for a textbook example.

Finally, what happens to your CCD if a pixel is saturated? Some charge may flow into adjacent pixels. Some digital cameras also don't actually feed you the raw information and you may be seeing some processing artefacts.

 

[LuckyFox]

This is almost surely a detector artifact of one sort or another.

Detector artefact in the form of spots of approximately the same size, which are recorded say once a week, not one after another, but after an hour, and in the same area of matrix (the next group of events is in a different area)? What mechanism could provide such an artifact?

 

[LuckyFox]

You actually only need 24 events before you've got better than a 50/50 chance that two occur in the same region.

Almost every group of such events has hits in the same area.

First is how sure are you that these events are actually cosmic rays?

No, this is just the version that seemed most plausible to me. What other source could there be - the Earth's core?

 

[Ibix]

Almost every group of such events has hits in the same area.

How many events in what you are calling a "group"?

No, this is just the version that seemed most plausible to me. What other source could there be - the Earth's core?

Detector noise is the obvious one. As I noted, DECO says it uses thresholding to identify candidate images, but then "performs more thorough follow-up processing to determine if the candidate should be considered an event". You only seem to be doing the first part, so you could just be seeing noise. Have you tried at different altitudes, or shielded?

 

[LuckyFox]

Detector noise is the obvious one.

Detector noise in the form of big spots of approximately the same size, which are recorded say once a week, not one after another, but after an hour, and in the same area of matrix (the next group of events is in a different area)?

How many events in what you are calling a "group"?

Usually three (sometimes two or four) events occur in one episode.

 

[LuckyFox]

Some test of my detector - old airplane tachometer with radium phosphor:

1440 events per day !!! (usually - 10...20)

 

[LuckyFox]

Differences between regular events and "big spots":

 

[Vanadium 50]

First, you haven't convinced me you are looking at cosmic rays at all. No matter what the app title says.

The cosmic ray rate is one through your head every second: 1 kHz/m². Your sensor, therefore, at maybe half a square centimeter would get one every 20 seconds. If there are 5 million pixels, each pizel would get a hit every 100 million seconds. Every three years.

No detector is that quiet that you get one noise pulse every three years. Conclusion - your "cosmic rays" have a lot of noise in them. Maybe predominantly noise.

The second point is that seeing two events in the exact same pixel two hours apart is evidence against cosmic rays. Your phone is sitting on a rotating earth. It moved 2000 miles between hits. How can two hits from the same source end up in the same pixel if the phone has moved 2000 miles?

 

[LuckyFox]

First, you haven't convinced me you are looking at cosmic rays at all. No matter what the app title says.

Cosmic particles are just a version; it’s not for nothing that I put question marks there. But I am very curious, what mechanism is responsible for such effects?

one noise pulse

Here we go again :-) It's not a single pixel, these are groups of pixels that form circle-like shapes, approximately the same size and the same brightness. Don't you think that the manifestation of such an intricate noise in the form of three circles located next to each other in one episode, and two overlapping circles in the next episode, although these large spots themselves are rare, is a very unlikely event?

 

[Ibix]

It's not a single pixel, these are groups of pixels that form circle-like shapes

You didn't answer when I asked how your detector reacts to a pixel being saturated. Nor about how certain you are that you are getting raw data from the camera. Some don't have completely uniform pixel grids and do some pre-processing to disguise that in normal "photo of your family" type images, which these aren't. That could well produce unusual effects.

Don't you think that the manifestation of such an intricate noise in the form of three circles located next to each other in one episode, and two overlapping circles in the next episode, although these large spots themselves are rare, is a very unlikely event?

As @Vanadium 50 pointed out, the spots are hours apart and the Earth turns - those spots are hundreds of kilometres apart at least. If those coincidences are above background (and you don't appear to have done any statistical analysis, just shown us a couple of images) then this points to a local issue, most probably in the detector. That's why we keep coming back to noise.

 

[sophiecentaur]

This is almost surely a detector artefact of one sort or another.

I'm inclined to agree. The detector will be omnidirectional as there's no 'optics' involved so the position of dots can only be governed by the time of arrival of a particle with respect to the scanning system. Interpreting what you have detected is a subtle business. To get directivity, you need additional equipment.

To identify spurious local effects (or to eliminate the possibility that what you see is actually cosmic rays), the apparatus would need to be run in a 'lead' box to identify local interference sources. Dunno what sort of thickness would be necessary but lead sheet used by roofers is available in various thicknesses and can be rolled and shaped easily. I'm not too sure that lead would be the best material because of the secondary neutrons produced from heavy nuclei collisions. Hydrogen atoms are most effective at shielding because secondary neutrons are not produced by collisions. That would need a bit of research to find a cheap solution - perhaps a tank of water would be a good screen OR the screening in space craft is often organic materials. There's a lot written about shielding in space so get hunting.

Counting white pixels with and without the shield would give you the equivalent of a 'background count' for conventional radioactivity measurements.

Once you are sure that the white pixels are genuine cosmic rays, you could use the shielding box with the lid removed to give some directivity. That, combined with the diurnal rotation of the Earth could identify the Sun as a source or (taking a longer interval) to identify deep space origin. In short, you could have a cosmic ray telescope.

 

[Vanadium 50]

There is no way a "lead box" is going to work.

1. Cosmic rays are mostly muons and penetrate lead, especially it is thin enough to make a box.
2. Some of the backgrounds are internal to the phone, and some may be external.
3. The sensor is way, way too small to measure cosmic rays.

As far as finding a cosmic ray source, like the sun, not going to happen. The sun blocks cosmic rays, instead of producing them, and a single CCD measures position and not direction. Besides, lets work the numbers out. The sun covers about 1/80000the of the sky, which I will call 1/86400 for simplicity. That means that even with perfect directionality - and you don't have any - you get one event in 20 days. For the sun to be the darkest spot in the sky, you need to run ~5x as long, or 100 days.

Even with perfect circumstances and a magic detector, this would take months. In real life, the sensor on a camera is way, way, way too small to make this work.

I have no idea what this is - probably what pops out of the (substantial) processing when you get noise (Johnson, shot, and who knows what else), but it is absolutely not cosmic rays. You don't need to know what breed of horse it is to know it isn't a unicorn.

 

[sophiecentaur]

Cosmic rays are mostly muons and penetrate lead, especially it is thin enough to make a box

I wondered about that. You do need a lot around reactors.

internal to the phone,

Yep; no directionality at all.. All the chips are probably equally sensitive, I guess. It's just that the sensor chips have amplified outputs and low threshold energy. And there will be more non-light sensing chips than specific light sensing junctions. But individual cosmic particles must have plenty of energy to spare for any CMOS device.

The acid test would be to disconnect the sensor array ??

 

[Vanadium 50]

The median cosmic ray energy is about 1 GeV, which means it will penetrate 2 feet of steel. Lead? It's denser, so maybe 16-18 inches. And that buys you a factor of 2. If you need to get away from cosmic rays, you need to go down in a mine. Even there it never zeroes out completely.

But it's clear that whatever the OP is seeing, it is not cosmic rays. Numbers matter. And the numbers don't work out here.

 

[Rive]

This is almost surely a detector artifact of one sort or another.

I think it'll be the 'another'. That detector has no selectivity at all, and since it's not anything specially built for the role (shielding, noise control, meticulous care about any possibly unstable materials and isotopes in structural parts, cleaned for external contamination and sealed in a controlled environment... ).

My wild guess is that the small dots are electronic noise and possibly 'weak' local events/captures (alpha, beta) within the device: while the fat ones are close gamma events.

Radon daughters (for example) might be a cheap explanation for some of the locality.

If the whole cam goes bright once, or misses some frames entirely - now, that could be some sure-cosmic event.